This proposal describes a 5-year training program for the development of an academic career in Molecular Cancer Epidemiology. The Principal Investigator has just completed a PhD in Epidemiology and Public Health at Yale University, supplementing the MD training she completed in the 1990s. In the context of a unique inter-departmental collaboration between the Pathology and Epidemiology departments at Yale, her dissertation research focused on characterizing the differential expression of selected cancer-related proteins using immunofluorescence-based immunohistochemical methods on a cohort of 192 primary melanoma patients, on executing multivariate statistical algorithms to develop a multi-marker prognostic model and validating the model in a separate cohort of 246 cases of primary melanoma. She now seeks to expand her capabilities as a molecular cancer epidemiologist by 1) acquiring DNA-focused laboratory skills for assaying levels of germline and tumor-based somatic genetic variation 2) expanding her familiarity with multivariate statistical methods useful for building integrated prognostic models and 3) gaining additional experience with epidemiologic field work for collecting survey data and for handling and archiving biospecimens. This program will continue the Principal Investigator's unique cross-disciplinary training in melanoma prognosis. Dr. David L. Rimm, Professor of Pathology and Director of Yale Pathology Tissue Services, and Dr. Susan T. Mayne, Professor of Epidemiology and Public Health, Associate Director for Population Sciences at the Yale Cancer Center and Program Leader for the Cancer Prevention and Control Research Program will jointly co-mentor the Principal Investigator's scientific development. Dr. Rimm is a recognized leader in the field of tissue microarrays and cancer biomarkers and is the inventor of Automated Quantitative Analysis (AQUA(R)) for characterizing levels of protein expression using immunohistochemistry. Dr. Mayne has directed cancer epidemiology studies since 1987, including case-control or case-based studies of lung, head and neck, breast, esophageal, gastric, pancreatic and endometrial cancers. Dr. Robert L. Camp, an established bioinformatician, will provide additional guidance during the statistical analysis phase. Research will focus on developing a molecularly-based multi-marker prognostic model for patients with newly-diagnosed metastatic melanoma. Metastatic melanoma has an overall poor prognosis with a median survival of <12 months. Yet, 25% of individuals do survive 2 or more years and ~5% will survive 5 or more years following the clinical diagnosis of metastatic disease. The ability to discriminate among subgroups of metastatic melanoma patients could optimize patient care in the clinic. Unchanged over the past 25 years, only a few clinicopathologic variables presently serve as validated prognostic factors and best multivariate models are associated with substantial prediction error for outcome. The objective of this proposal is to generate and validate a prognostic model for metastatic melanoma that integrates clinicopathologic, genetic, and expression data, utilizing two independent, well-annotated cohorts of patients who underwent metastatic melanoma resections at Yale-New Haven Hospital. The Specific Aims include: 1) to build an integrated prognostic model for time to death from metastatic melanoma from clinicopathologic variables, germline genetic variation, as well as tumor-based somatic mutations and protein expression for selected candidate genes of known relevance to melanoma biology in a retrospective Discovery cohort of 217 patients who underwent a metastatic melanoma resection at Yale-New Haven Hospital during 1959-1994 and 2) to validate the metastatic melanoma prognostic model in an independent retrospective cohort of patients with metastatic melanoma resected at Yale-New Haven Hospital during 1995-2008. Proposed experiments include the assembly and clinical annotation of both the Yale Melanoma Metastasis Discovery (1959-1994) and Validation (1995-2008) cohorts, the performance of wet-bench experiments to collect exposure information for selected molecular candidates and the execution of basic survival analyses as well as higher-order multivariate statistical methods to construct a multi-marker prognostic model for metastatic melanoma. In addition to the expression profiles for 80 candidate proteins and B-RAF and N-RAS somatic mutations previously collected on the Discovery cohort, this proposal will support work to characterize differential expression for 25 additional proteins, somatic mutations for 10 genes and germline genetic variation for 10 genes. Candidates will be selected from published data demonstrating relevance to melanoma metastasis. This proposal also plans to evaluate the association of lifestyle choices such as body mass index, and tobacco and alcohol use with survival in this population. This proposal will be, to the best of our knowledge, the first attempt to develop a REMARK-compliant prognostic model for metastatic melanoma that incorporates molecular exposures. The Pathology Department at Yale University provides an ideal setting for the proposed research. In addition to access to state-of-the-art facilities and equipment, she will also benefit from interactions with prominent junior and senior cancer investigators among the faculty. The PI will also participate in the Institution's Melanoma SPORE research program (Ruth Halaban, PI) and benefit from Yale's institutional commitment to personalized cancer therapy, supported by its hiring of Dr. Thomas Lynch as Physician-in-Chief of the new cancer hospital.